DarkKitsune/Qwen3.8-Flash-Next-UD-IQ3_XXS-PLE-25pct
Quantization vs unsloth/Qwen3.8-Flash-Next-GGUF. DarkKitsune/Qwen3.8-Flash-Next-UD-IQ3_XXS-PLE-25pct is a GGUF quantization (IQ4_NL/IQ2_S) of the same weights: 581 tensors match exactly and 642 within quantization error.
Tensors changed: Norms 0/184, other 0/222, Attention 0/288, MLP 0/528, lm_head 0/1